read-cache.c

   1 /*
   2  * GIT - The information manager from hell
   3  *
   4  * Copyright (C) Linus Torvalds, 2005
   5  */
   6 #define NO_THE_INDEX_COMPATIBILITY_MACROS
   7 #include "cache.h"
   8 #include "cache-tree.h"
   9 #include "refs.h"
  10 #include "dir.h"
  11
  12 /* Index extensions.
  13  *
  14  * The first letter should be 'A'..'Z' for extensions that are not
  15  * necessary for a correct operation (i.e. optimization data).
  16  * When new extensions are added that _needs_ to be understood in
  17  * order to correctly interpret the index file, pick character that
  18  * is outside the range, to cause the reader to abort.
  19  */
  20
  21 #define CACHE_EXT(s) ( (s[0]<<24)|(s[1]<<16)|(s[2]<<8)|(s[3]) )
  22 #define CACHE_EXT_TREE 0x54524545       /* "TREE" */
  23
  24 struct index_state the_index;
  25
  26 /*
  27  * This only updates the "non-critical" parts of the directory
  28  * cache, ie the parts that aren't tracked by GIT, and only used
  29  * to validate the cache.
  30  */
  31 void fill_stat_cache_info(struct cache_entry *ce, struct stat *st)
  32 {
  33         ce->ce_ctime.sec = htonl(st->st_ctime);
  34         ce->ce_mtime.sec = htonl(st->st_mtime);
  35 #ifdef USE_NSEC
  36         ce->ce_ctime.nsec = htonl(st->st_ctim.tv_nsec);
  37         ce->ce_mtime.nsec = htonl(st->st_mtim.tv_nsec);
  38 #endif
  39         ce->ce_dev = htonl(st->st_dev);
  40         ce->ce_ino = htonl(st->st_ino);
  41         ce->ce_uid = htonl(st->st_uid);
  42         ce->ce_gid = htonl(st->st_gid);
  43         ce->ce_size = htonl(st->st_size);
  44
  45         if (assume_unchanged)
  46                 ce->ce_flags |= htons(CE_VALID);
  47 }
  48
  49 static int ce_compare_data(struct cache_entry *ce, struct stat *st)
  50 {
  51         int match = -1;
  52         int fd = open(ce->name, O_RDONLY);
  53
  54         if (fd >= 0) {
  55                 unsigned char sha1[20];
  56                 if (!index_fd(sha1, fd, st, 0, OBJ_BLOB, ce->name))
  57                         match = hashcmp(sha1, ce->sha1);
  58                 /* index_fd() closed the file descriptor already */
  59         }
  60         return match;
  61 }
  62
  63 static int ce_compare_link(struct cache_entry *ce, size_t expected_size)
  64 {
  65         int match = -1;
  66         char *target;
  67         void *buffer;
  68         unsigned long size;
  69         enum object_type type;
  70         int len;
  71
  72         target = xmalloc(expected_size);
  73         len = readlink(ce->name, target, expected_size);
  74         if (len != expected_size) {
  75                 free(target);
  76                 return -1;
  77         }
  78         buffer = read_sha1_file(ce->sha1, &type, &size);
  79         if (!buffer) {
  80                 free(target);
  81                 return -1;
  82         }
  83         if (size == expected_size)
  84                 match = memcmp(buffer, target, size);
  85         free(buffer);
  86         free(target);
  87         return match;
  88 }
  89
  90 static int ce_compare_gitlink(struct cache_entry *ce)
  91 {
  92         unsigned char sha1[20];
  93
  94         /*
  95          * We don't actually require that the .git directory
  96          * under GITLINK directory be a valid git directory. It
  97          * might even be missing (in case nobody populated that
  98          * sub-project).
  99          *
 100          * If so, we consider it always to match.
 101          */
 102         if (resolve_gitlink_ref(ce->name, "HEAD", sha1) < 0)
 103                 return 0;
 104         return hashcmp(sha1, ce->sha1);
 105 }
 106
 107 static int ce_modified_check_fs(struct cache_entry *ce, struct stat *st)
 108 {
 109         switch (st->st_mode & S_IFMT) {
 110         case S_IFREG:
 111                 if (ce_compare_data(ce, st))
 112                         return DATA_CHANGED;
 113                 break;
 114         case S_IFLNK:
 115                 if (ce_compare_link(ce, xsize_t(st->st_size)))
 116                         return DATA_CHANGED;
 117                 break;
 118         case S_IFDIR:
 119                 if (S_ISGITLINK(ntohl(ce->ce_mode)))
 120                         return 0;
 121         default:
 122                 return TYPE_CHANGED;
 123         }
 124         return 0;
 125 }
 126
 127 static int ce_match_stat_basic(struct cache_entry *ce, struct stat *st)
 128 {
 129         unsigned int changed = 0;
 130
 131         switch (ntohl(ce->ce_mode) & S_IFMT) {
 132         case S_IFREG:
 133                 changed |= !S_ISREG(st->st_mode) ? TYPE_CHANGED : 0;
 134                 /* We consider only the owner x bit to be relevant for
 135                  * "mode changes"
 136                  */
 137                 if (trust_executable_bit &&
 138                     (0100 & (ntohl(ce->ce_mode) ^ st->st_mode)))
 139                         changed |= MODE_CHANGED;
 140                 break;
 141         case S_IFLNK:
 142                 if (!S_ISLNK(st->st_mode) &&
 143                     (has_symlinks || !S_ISREG(st->st_mode)))
 144                         changed |= TYPE_CHANGED;
 145                 break;
 146         case S_IFGITLINK:
 147                 if (!S_ISDIR(st->st_mode))
 148                         changed |= TYPE_CHANGED;
 149                 else if (ce_compare_gitlink(ce))
 150                         changed |= DATA_CHANGED;
 151                 return changed;
 152         case 0: /* Special case: unmerged file in index */
 153                 return MODE_CHANGED | DATA_CHANGED | TYPE_CHANGED;
 154         default:
 155                 die("internal error: ce_mode is %o", ntohl(ce->ce_mode));
 156         }
 157         if (ce->ce_mtime.sec != htonl(st->st_mtime))
 158                 changed |= MTIME_CHANGED;
 159         if (ce->ce_ctime.sec != htonl(st->st_ctime))
 160                 changed |= CTIME_CHANGED;
 161
 162 #ifdef USE_NSEC
 163         /*
 164          * nsec seems unreliable - not all filesystems support it, so
 165          * as long as it is in the inode cache you get right nsec
 166          * but after it gets flushed, you get zero nsec.
 167          */
 168         if (ce->ce_mtime.nsec != htonl(st->st_mtim.tv_nsec))
 169                 changed |= MTIME_CHANGED;
 170         if (ce->ce_ctime.nsec != htonl(st->st_ctim.tv_nsec))
 171                 changed |= CTIME_CHANGED;
 172 #endif
 173
 174         if (ce->ce_uid != htonl(st->st_uid) ||
 175             ce->ce_gid != htonl(st->st_gid))
 176                 changed |= OWNER_CHANGED;
 177         if (ce->ce_ino != htonl(st->st_ino))
 178                 changed |= INODE_CHANGED;
 179
 180 #ifdef USE_STDEV
 181         /*
 182          * st_dev breaks on network filesystems where different
 183          * clients will have different views of what "device"
 184          * the filesystem is on
 185          */
 186         if (ce->ce_dev != htonl(st->st_dev))
 187                 changed |= INODE_CHANGED;
 188 #endif
 189
 190         if (ce->ce_size != htonl(st->st_size))
 191                 changed |= DATA_CHANGED;
 192
 193         return changed;
 194 }
 195
 196 int ie_match_stat(struct index_state *istate,
 197                   struct cache_entry *ce, struct stat *st, int options)
 198 {
 199         unsigned int changed;
 200         int ignore_valid = options & 01;
 201         int assume_racy_is_modified = options & 02;
 202
 203         /*
 204          * If it's marked as always valid in the index, it's
 205          * valid whatever the checked-out copy says.
 206          */
 207         if (!ignore_valid && (ce->ce_flags & htons(CE_VALID)))
 208                 return 0;
 209
 210         changed = ce_match_stat_basic(ce, st);
 211
 212         /*
 213          * Within 1 second of this sequence:
 214          *      echo xyzzy >file && git-update-index --add file
 215          * running this command:
 216          *      echo frotz >file
 217          * would give a falsely clean cache entry.  The mtime and
 218          * length match the cache, and other stat fields do not change.
 219          *
 220          * We could detect this at update-index time (the cache entry
 221          * being registered/updated records the same time as "now")
 222          * and delay the return from git-update-index, but that would
 223          * effectively mean we can make at most one commit per second,
 224          * which is not acceptable.  Instead, we check cache entries
 225          * whose mtime are the same as the index file timestamp more
 226          * carefully than others.
 227          */
 228         if (!changed &&
 229             istate->timestamp &&
 230             istate->timestamp <= ntohl(ce->ce_mtime.sec)) {
 231                 if (assume_racy_is_modified)
 232                         changed |= DATA_CHANGED;
 233                 else
 234                         changed |= ce_modified_check_fs(ce, st);
 235         }
 236
 237         return changed;
 238 }
 239
 240 int ie_modified(struct index_state *istate,
 241                 struct cache_entry *ce, struct stat *st, int really)
 242 {
 243         int changed, changed_fs;
 244         changed = ie_match_stat(istate, ce, st, really);
 245         if (!changed)
 246                 return 0;
 247         /*
 248          * If the mode or type has changed, there's no point in trying
 249          * to refresh the entry - it's not going to match
 250          */
 251         if (changed & (MODE_CHANGED | TYPE_CHANGED))
 252                 return changed;
 253
 254         /* Immediately after read-tree or update-index --cacheinfo,
 255          * the length field is zero.  For other cases the ce_size
 256          * should match the SHA1 recorded in the index entry.
 257          */
 258         if ((changed & DATA_CHANGED) && ce->ce_size != htonl(0))
 259                 return changed;
 260
 261         changed_fs = ce_modified_check_fs(ce, st);
 262         if (changed_fs)
 263                 return changed | changed_fs;
 264         return 0;
 265 }
 266
 267 int base_name_compare(const char *name1, int len1, int mode1,
 268                       const char *name2, int len2, int mode2)
 269 {
 270         unsigned char c1, c2;
 271         int len = len1 < len2 ? len1 : len2;
 272         int cmp;
 273
 274         cmp = memcmp(name1, name2, len);
 275         if (cmp)
 276                 return cmp;
 277         c1 = name1[len];
 278         c2 = name2[len];
 279         if (!c1 && S_ISDIR(mode1))
 280                 c1 = '/';
 281         if (!c2 && S_ISDIR(mode2))
 282                 c2 = '/';
 283         return (c1 < c2) ? -1 : (c1 > c2) ? 1 : 0;
 284 }
 285
 286 int cache_name_compare(const char *name1, int flags1, const char *name2, int flags2)
 287 {
 288         int len1 = flags1 & CE_NAMEMASK;
 289         int len2 = flags2 & CE_NAMEMASK;
 290         int len = len1 < len2 ? len1 : len2;
 291         int cmp;
 292
 293         cmp = memcmp(name1, name2, len);
 294         if (cmp)
 295                 return cmp;
 296         if (len1 < len2)
 297                 return -1;
 298         if (len1 > len2)
 299                 return 1;
 300
 301         /* Compare stages  */
 302         flags1 &= CE_STAGEMASK;
 303         flags2 &= CE_STAGEMASK;
 304
 305         if (flags1 < flags2)
 306                 return -1;
 307         if (flags1 > flags2)
 308                 return 1;
 309         return 0;
 310 }
 311
 312 int index_name_pos(struct index_state *istate, const char *name, int namelen)
 313 {
 314         int first, last;
 315
 316         first = 0;
 317         last = istate->cache_nr;
 318         while (last > first) {
 319                 int next = (last + first) >> 1;
 320                 struct cache_entry *ce = istate->cache[next];
 321                 int cmp = cache_name_compare(name, namelen, ce->name, ntohs(ce->ce_flags));
 322                 if (!cmp)
 323                         return next;
 324                 if (cmp < 0) {
 325                         last = next;
 326                         continue;
 327                 }
 328                 first = next+1;
 329         }
 330         return -first-1;
 331 }
 332
 333 /* Remove entry, return true if there are more entries to go.. */
 334 int remove_index_entry_at(struct index_state *istate, int pos)
 335 {
 336         istate->cache_changed = 1;
 337         istate->cache_nr--;
 338         if (pos >= istate->cache_nr)
 339                 return 0;
 340         memmove(istate->cache + pos,
 341                 istate->cache + pos + 1,
 342                 (istate->cache_nr - pos) * sizeof(struct cache_entry *));
 343         return 1;
 344 }
 345
 346 int remove_file_from_index(struct index_state *istate, const char *path)
 347 {
 348         int pos = index_name_pos(istate, path, strlen(path));
 349         if (pos < 0)
 350                 pos = -pos-1;
 351         while (pos < istate->cache_nr && !strcmp(istate->cache[pos]->name, path))
 352                 remove_index_entry_at(istate, pos);
 353         return 0;
 354 }
 355
 356 static int compare_name(struct cache_entry *ce, const char *path, int namelen)
 357 {
 358         return namelen != ce_namelen(ce) || memcmp(path, ce->name, namelen);
 359 }
 360
 361 static int index_name_pos_also_unmerged(struct index_state *istate,
 362         const char *path, int namelen)
 363 {
 364         int pos = index_name_pos(istate, path, namelen);
 365         struct cache_entry *ce;
 366
 367         if (pos >= 0)
 368                 return pos;
 369
 370         /* maybe unmerged? */
 371         pos = -1 - pos;
 372         if (pos >= istate->cache_nr ||
 373                         compare_name((ce = istate->cache[pos]), path, namelen))
 374                 return -1;
 375
 376         /* order of preference: stage 2, 1, 3 */
 377         if (ce_stage(ce) == 1 && pos + 1 < istate->cache_nr &&
 378                         ce_stage((ce = istate->cache[pos + 1])) == 2 &&
 379                         !compare_name(ce, path, namelen))
 380                 pos++;
 381         return pos;
 382 }
 383
 384 int add_file_to_index(struct index_state *istate, const char *path, int verbose)
 385 {
 386         int size, namelen, pos;
 387         struct stat st;
 388         struct cache_entry *ce;
 389
 390         if (lstat(path, &st))
 391                 die("%s: unable to stat (%s)", path, strerror(errno));
 392
 393         if (!S_ISREG(st.st_mode) && !S_ISLNK(st.st_mode) && !S_ISDIR(st.st_mode))
 394                 die("%s: can only add regular files, symbolic links or git-directories", path);
 395
 396         namelen = strlen(path);
 397         if (S_ISDIR(st.st_mode)) {
 398                 while (namelen && path[namelen-1] == '/')
 399                         namelen--;
 400         }
 401         size = cache_entry_size(namelen);
 402         ce = xcalloc(1, size);
 403         memcpy(ce->name, path, namelen);
 404         ce->ce_flags = htons(namelen);
 405         fill_stat_cache_info(ce, &st);
 406
 407         if (trust_executable_bit && has_symlinks)
 408                 ce->ce_mode = create_ce_mode(st.st_mode);
 409         else {
 410                 /* If there is an existing entry, pick the mode bits and type
 411                  * from it, otherwise assume unexecutable regular file.
 412                  */
 413                 struct cache_entry *ent;
 414                 int pos = index_name_pos_also_unmerged(istate, path, namelen);
 415
 416                 ent = (0 <= pos) ? istate->cache[pos] : NULL;
 417                 ce->ce_mode = ce_mode_from_stat(ent, st.st_mode);
 418         }
 419
 420         pos = index_name_pos(istate, ce->name, namelen);
 421         if (0 <= pos &&
 422             !ce_stage(istate->cache[pos]) &&
 423             !ie_modified(istate, istate->cache[pos], &st, 1)) {
 424                 /* Nothing changed, really */
 425                 free(ce);
 426                 return 0;
 427         }
 428
 429         if (index_path(ce->sha1, path, &st, 1))
 430                 die("unable to index file %s", path);
 431         if (add_index_entry(istate, ce, ADD_CACHE_OK_TO_ADD|ADD_CACHE_OK_TO_REPLACE))
 432                 die("unable to add %s to index",path);
 433         if (verbose)
 434                 printf("add '%s'\n", path);
 435         cache_tree_invalidate_path(istate->cache_tree, path);
 436         return 0;
 437 }
 438
 439 struct cache_entry *make_cache_entry(unsigned int mode,
 440                 const unsigned char *sha1, const char *path, int stage,
 441                 int refresh)
 442 {
 443         int size, len;
 444         struct cache_entry *ce;
 445
 446         if (!verify_path(path))
 447                 return NULL;
 448
 449         len = strlen(path);
 450         size = cache_entry_size(len);
 451         ce = xcalloc(1, size);
 452
 453         hashcpy(ce->sha1, sha1);
 454         memcpy(ce->name, path, len);
 455         ce->ce_flags = create_ce_flags(len, stage);
 456         ce->ce_mode = create_ce_mode(mode);
 457
 458         if (refresh)
 459                 return refresh_cache_entry(ce, 0);
 460
 461         return ce;
 462 }
 463
 464 int ce_same_name(struct cache_entry *a, struct cache_entry *b)
 465 {
 466         int len = ce_namelen(a);
 467         return ce_namelen(b) == len && !memcmp(a->name, b->name, len);
 468 }
 469
 470 int ce_path_match(const struct cache_entry *ce, const char **pathspec)
 471 {
 472         const char *match, *name;
 473         int len;
 474
 475         if (!pathspec)
 476                 return 1;
 477
 478         len = ce_namelen(ce);
 479         name = ce->name;
 480         while ((match = *pathspec++) != NULL) {
 481                 int matchlen = strlen(match);
 482                 if (matchlen > len)
 483                         continue;
 484                 if (memcmp(name, match, matchlen))
 485                         continue;
 486                 if (matchlen && name[matchlen-1] == '/')
 487                         return 1;
 488                 if (name[matchlen] == '/' || !name[matchlen])
 489                         return 1;
 490                 if (!matchlen)
 491                         return 1;
 492         }
 493         return 0;
 494 }
 495
 496 /*
 497  * We fundamentally don't like some paths: we don't want
 498  * dot or dot-dot anywhere, and for obvious reasons don't
 499  * want to recurse into ".git" either.
 500  *
 501  * Also, we don't want double slashes or slashes at the
 502  * end that can make pathnames ambiguous.
 503  */
 504 static int verify_dotfile(const char *rest)
 505 {
 506         /*
 507          * The first character was '.', but that
 508          * has already been discarded, we now test
 509          * the rest.
 510          */
 511         switch (*rest) {
 512         /* "." is not allowed */
 513         case '\0': case '/':
 514                 return 0;
 515
 516         /*
 517          * ".git" followed by  NUL or slash is bad. This
 518          * shares the path end test with the ".." case.
 519          */
 520         case 'g':
 521                 if (rest[1] != 'i')
 522                         break;
 523                 if (rest[2] != 't')
 524                         break;
 525                 rest += 2;
 526         /* fallthrough */
 527         case '.':
 528                 if (rest[1] == '\0' || rest[1] == '/')
 529                         return 0;
 530         }
 531         return 1;
 532 }
 533
 534 int verify_path(const char *path)
 535 {
 536         char c;
 537
 538         goto inside;
 539         for (;;) {
 540                 if (!c)
 541                         return 1;
 542                 if (c == '/') {
 543 inside:
 544                         c = *path++;
 545                         switch (c) {
 546                         default:
 547                                 continue;
 548                         case '/': case '\0':
 549                                 break;
 550                         case '.':
 551                                 if (verify_dotfile(path))
 552                                         continue;
 553                         }
 554                         return 0;
 555                 }
 556                 c = *path++;
 557         }
 558 }
 559
 560 /*
 561  * Do we have another file that has the beginning components being a
 562  * proper superset of the name we're trying to add?
 563  */
 564 static int has_file_name(struct index_state *istate,
 565                          const struct cache_entry *ce, int pos, int ok_to_replace)
 566 {
 567         int retval = 0;
 568         int len = ce_namelen(ce);
 569         int stage = ce_stage(ce);
 570         const char *name = ce->name;
 571
 572         while (pos < istate->cache_nr) {
 573                 struct cache_entry *p = istate->cache[pos++];
 574
 575                 if (len >= ce_namelen(p))
 576                         break;
 577                 if (memcmp(name, p->name, len))
 578                         break;
 579                 if (ce_stage(p) != stage)
 580                         continue;
 581                 if (p->name[len] != '/')
 582                         continue;
 583                 if (!ce_stage(p) && !p->ce_mode)
 584                         continue;
 585                 retval = -1;
 586                 if (!ok_to_replace)
 587                         break;
 588                 remove_index_entry_at(istate, --pos);
 589         }
 590         return retval;
 591 }
 592
 593 /*
 594  * Do we have another file with a pathname that is a proper
 595  * subset of the name we're trying to add?
 596  */
 597 static int has_dir_name(struct index_state *istate,
 598                         const struct cache_entry *ce, int pos, int ok_to_replace)
 599 {
 600         int retval = 0;
 601         int stage = ce_stage(ce);
 602         const char *name = ce->name;
 603         const char *slash = name + ce_namelen(ce);
 604
 605         for (;;) {
 606                 int len;
 607
 608                 for (;;) {
 609                         if (*--slash == '/')
 610                                 break;
 611                         if (slash <= ce->name)
 612                                 return retval;
 613                 }
 614                 len = slash - name;
 615
 616                 pos = index_name_pos(istate, name, ntohs(create_ce_flags(len, stage)));
 617                 if (pos >= 0) {
 618                         /*
 619                          * Found one, but not so fast.  This could
 620                          * be a marker that says "I was here, but
 621                          * I am being removed".  Such an entry is
 622                          * not a part of the resulting tree, and
 623                          * it is Ok to have a directory at the same
 624                          * path.
 625                          */
 626                         if (stage || istate->cache[pos]->ce_mode) {
 627                                 retval = -1;
 628                                 if (!ok_to_replace)
 629                                         break;
 630                                 remove_index_entry_at(istate, pos);
 631                                 continue;
 632                         }
 633                 }
 634                 else
 635                         pos = -pos-1;
 636
 637                 /*
 638                  * Trivial optimization: if we find an entry that
 639                  * already matches the sub-directory, then we know
 640                  * we're ok, and we can exit.
 641                  */
 642                 while (pos < istate->cache_nr) {
 643                         struct cache_entry *p = istate->cache[pos];
 644                         if ((ce_namelen(p) <= len) ||
 645                             (p->name[len] != '/') ||
 646                             memcmp(p->name, name, len))
 647                                 break; /* not our subdirectory */
 648                         if (ce_stage(p) == stage && (stage || p->ce_mode))
 649                                 /* p is at the same stage as our entry, and
 650                                  * is a subdirectory of what we are looking
 651                                  * at, so we cannot have conflicts at our
 652                                  * level or anything shorter.
 653                                  */
 654                                 return retval;
 655                         pos++;
 656                 }
 657         }
 658         return retval;
 659 }
 660
 661 /* We may be in a situation where we already have path/file and path
 662  * is being added, or we already have path and path/file is being
 663  * added.  Either one would result in a nonsense tree that has path
 664  * twice when git-write-tree tries to write it out.  Prevent it.
 665  *
 666  * If ok-to-replace is specified, we remove the conflicting entries
 667  * from the cache so the caller should recompute the insert position.
 668  * When this happens, we return non-zero.
 669  */
 670 static int check_file_directory_conflict(struct index_state *istate,
 671                                          const struct cache_entry *ce,
 672                                          int pos, int ok_to_replace)
 673 {
 674         int retval;
 675
 676         /*
 677          * When ce is an "I am going away" entry, we allow it to be added
 678          */
 679         if (!ce_stage(ce) && !ce->ce_mode)
 680                 return 0;
 681
 682         /*
 683          * We check if the path is a sub-path of a subsequent pathname
 684          * first, since removing those will not change the position
 685          * in the array.
 686          */
 687         retval = has_file_name(istate, ce, pos, ok_to_replace);
 688
 689         /*
 690          * Then check if the path might have a clashing sub-directory
 691          * before it.
 692          */
 693         return retval + has_dir_name(istate, ce, pos, ok_to_replace);
 694 }
 695
 696 static int add_index_entry_with_check(struct index_state *istate, struct cache_entry *ce, int option)
 697 {
 698         int pos;
 699         int ok_to_add = option & ADD_CACHE_OK_TO_ADD;
 700         int ok_to_replace = option & ADD_CACHE_OK_TO_REPLACE;
 701         int skip_df_check = option & ADD_CACHE_SKIP_DFCHECK;
 702
 703         pos = index_name_pos(istate, ce->name, ntohs(ce->ce_flags));
 704
 705         /* existing match? Just replace it. */
 706         if (pos >= 0) {
 707                 istate->cache_changed = 1;
 708                 istate->cache[pos] = ce;
 709                 return 0;
 710         }
 711         pos = -pos-1;
 712
 713         /*
 714          * Inserting a merged entry ("stage 0") into the index
 715          * will always replace all non-merged entries..
 716          */
 717         if (pos < istate->cache_nr && ce_stage(ce) == 0) {
 718                 while (ce_same_name(istate->cache[pos], ce)) {
 719                         ok_to_add = 1;
 720                         if (!remove_index_entry_at(istate, pos))
 721                                 break;
 722                 }
 723         }
 724
 725         if (!ok_to_add)
 726                 return -1;
 727         if (!verify_path(ce->name))
 728                 return -1;
 729
 730         if (!skip_df_check &&
 731             check_file_directory_conflict(istate, ce, pos, ok_to_replace)) {
 732                 if (!ok_to_replace)
 733                         return error("'%s' appears as both a file and as a directory",
 734                                      ce->name);
 735                 pos = index_name_pos(istate, ce->name, ntohs(ce->ce_flags));
 736                 pos = -pos-1;
 737         }
 738         return pos + 1;
 739 }
 740
 741 int add_index_entry(struct index_state *istate, struct cache_entry *ce, int option)
 742 {
 743         int pos;
 744
 745         if (option & ADD_CACHE_JUST_APPEND)
 746                 pos = istate->cache_nr;
 747         else {
 748                 int ret;
 749                 ret = add_index_entry_with_check(istate, ce, option);
 750                 if (ret <= 0)
 751                         return ret;
 752                 pos = ret - 1;
 753         }
 754
 755         /* Make sure the array is big enough .. */
 756         if (istate->cache_nr == istate->cache_alloc) {
 757                 istate->cache_alloc = alloc_nr(istate->cache_alloc);
 758                 istate->cache = xrealloc(istate->cache,
 759                                         istate->cache_alloc * sizeof(struct cache_entry *));
 760         }
 761
 762         /* Add it in.. */
 763         istate->cache_nr++;
 764         if (istate->cache_nr > pos + 1)
 765                 memmove(istate->cache + pos + 1,
 766                         istate->cache + pos,
 767                         (istate->cache_nr - pos - 1) * sizeof(ce));
 768         istate->cache[pos] = ce;
 769         istate->cache_changed = 1;
 770         return 0;
 771 }
 772
 773 /*
 774  * "refresh" does not calculate a new sha1 file or bring the
 775  * cache up-to-date for mode/content changes. But what it
 776  * _does_ do is to "re-match" the stat information of a file
 777  * with the cache, so that you can refresh the cache for a
 778  * file that hasn't been changed but where the stat entry is
 779  * out of date.
 780  *
 781  * For example, you'd want to do this after doing a "git-read-tree",
 782  * to link up the stat cache details with the proper files.
 783  */
 784 static struct cache_entry *refresh_cache_ent(struct index_state *istate,
 785                                              struct cache_entry *ce, int really, int *err)
 786 {
 787         struct stat st;
 788         struct cache_entry *updated;
 789         int changed, size;
 790
 791         if (lstat(ce->name, &st) < 0) {
 792                 if (err)
 793                         *err = errno;
 794                 return NULL;
 795         }
 796
 797         changed = ie_match_stat(istate, ce, &st, really);
 798         if (!changed) {
 799                 if (really && assume_unchanged &&
 800                     !(ce->ce_flags & htons(CE_VALID)))
 801                         ; /* mark this one VALID again */
 802                 else
 803                         return ce;
 804         }
 805
 806         if (ie_modified(istate, ce, &st, really)) {
 807                 if (err)
 808                         *err = EINVAL;
 809                 return NULL;
 810         }
 811
 812         size = ce_size(ce);
 813         updated = xmalloc(size);
 814         memcpy(updated, ce, size);
 815         fill_stat_cache_info(updated, &st);
 816
 817         /* In this case, if really is not set, we should leave
 818          * CE_VALID bit alone.  Otherwise, paths marked with
 819          * --no-assume-unchanged (i.e. things to be edited) will
 820          * reacquire CE_VALID bit automatically, which is not
 821          * really what we want.
 822          */
 823         if (!really && assume_unchanged && !(ce->ce_flags & htons(CE_VALID)))
 824                 updated->ce_flags &= ~htons(CE_VALID);
 825
 826         return updated;
 827 }
 828
 829 int refresh_index(struct index_state *istate, unsigned int flags, const char **pathspec, char *seen)
 830 {
 831         int i;
 832         int has_errors = 0;
 833         int really = (flags & REFRESH_REALLY) != 0;
 834         int allow_unmerged = (flags & REFRESH_UNMERGED) != 0;
 835         int quiet = (flags & REFRESH_QUIET) != 0;
 836         int not_new = (flags & REFRESH_IGNORE_MISSING) != 0;
 837
 838         for (i = 0; i < istate->cache_nr; i++) {
 839                 struct cache_entry *ce, *new;
 840                 int cache_errno = 0;
 841
 842                 ce = istate->cache[i];
 843                 if (ce_stage(ce)) {
 844                         while ((i < istate->cache_nr) &&
 845                                ! strcmp(istate->cache[i]->name, ce->name))
 846                                 i++;
 847                         i--;
 848                         if (allow_unmerged)
 849                                 continue;
 850                         printf("%s: needs merge\n", ce->name);
 851                         has_errors = 1;
 852                         continue;
 853                 }
 854
 855                 if (pathspec && !match_pathspec(pathspec, ce->name, strlen(ce->name), 0, seen))
 856                         continue;
 857
 858                 new = refresh_cache_ent(istate, ce, really, &cache_errno);
 859                 if (new == ce)
 860                         continue;
 861                 if (!new) {
 862                         if (not_new && cache_errno == ENOENT)
 863                                 continue;
 864                         if (really && cache_errno == EINVAL) {
 865                                 /* If we are doing --really-refresh that
 866                                  * means the index is not valid anymore.
 867                                  */
 868                                 ce->ce_flags &= ~htons(CE_VALID);
 869                                 istate->cache_changed = 1;
 870                         }
 871                         if (quiet)
 872                                 continue;
 873                         printf("%s: needs update\n", ce->name);
 874                         has_errors = 1;
 875                         continue;
 876                 }
 877                 istate->cache_changed = 1;
 878                 /* You can NOT just free istate->cache[i] here, since it
 879                  * might not be necessarily malloc()ed but can also come
 880                  * from mmap(). */
 881                 istate->cache[i] = new;
 882         }
 883         return has_errors;
 884 }
 885
 886 struct cache_entry *refresh_cache_entry(struct cache_entry *ce, int really)
 887 {
 888         return refresh_cache_ent(&the_index, ce, really, NULL);
 889 }
 890
 891 static int verify_hdr(struct cache_header *hdr, unsigned long size)
 892 {
 893         SHA_CTX c;
 894         unsigned char sha1[20];
 895
 896         if (hdr->hdr_signature != htonl(CACHE_SIGNATURE))
 897                 return error("bad signature");
 898         if (hdr->hdr_version != htonl(2))
 899                 return error("bad index version");
 900         SHA1_Init(&c);
 901         SHA1_Update(&c, hdr, size - 20);
 902         SHA1_Final(sha1, &c);
 903         if (hashcmp(sha1, (unsigned char *)hdr + size - 20))
 904                 return error("bad index file sha1 signature");
 905         return 0;
 906 }
 907
 908 static int read_index_extension(struct index_state *istate,
 909                                 const char *ext, void *data, unsigned long sz)
 910 {
 911         switch (CACHE_EXT(ext)) {
 912         case CACHE_EXT_TREE:
 913                 istate->cache_tree = cache_tree_read(data, sz);
 914                 break;
 915         default:
 916                 if (*ext < 'A' || 'Z' < *ext)
 917                         return error("index uses %.4s extension, which we do not understand",
 918                                      ext);
 919                 fprintf(stderr, "ignoring %.4s extension\n", ext);
 920                 break;
 921         }
 922         return 0;
 923 }
 924
 925 int read_index(struct index_state *istate)
 926 {
 927         return read_index_from(istate, get_index_file());
 928 }
 929
 930 /* remember to discard_cache() before reading a different cache! */
 931 int read_index_from(struct index_state *istate, const char *path)
 932 {
 933         int fd, i;
 934         struct stat st;
 935         unsigned long offset;
 936         struct cache_header *hdr;
 937
 938         errno = EBUSY;
 939         if (istate->mmap)
 940                 return istate->cache_nr;
 941
 942         errno = ENOENT;
 943         istate->timestamp = 0;
 944         fd = open(path, O_RDONLY);
 945         if (fd < 0) {
 946                 if (errno == ENOENT)
 947                         return 0;
 948                 die("index file open failed (%s)", strerror(errno));
 949         }
 950
 951         if (fstat(fd, &st))
 952                 die("cannot stat the open index (%s)", strerror(errno));
 953
 954         errno = EINVAL;
 955         istate->mmap_size = xsize_t(st.st_size);
 956         if (istate->mmap_size < sizeof(struct cache_header) + 20)
 957                 die("index file smaller than expected");
 958
 959         istate->mmap = xmmap(NULL, istate->mmap_size, PROT_READ | PROT_WRITE, MAP_PRIVATE, fd, 0);
 960         close(fd);
 961
 962         hdr = istate->mmap;
 963         if (verify_hdr(hdr, istate->mmap_size) < 0)
 964                 goto unmap;
 965
 966         istate->cache_nr = ntohl(hdr->hdr_entries);
 967         istate->cache_alloc = alloc_nr(istate->cache_nr);
 968         istate->cache = xcalloc(istate->cache_alloc, sizeof(struct cache_entry *));
 969
 970         offset = sizeof(*hdr);
 971         for (i = 0; i < istate->cache_nr; i++) {
 972                 struct cache_entry *ce;
 973
 974                 ce = (struct cache_entry *)((char *)(istate->mmap) + offset);
 975                 offset = offset + ce_size(ce);
 976                 istate->cache[i] = ce;
 977         }
 978         istate->timestamp = st.st_mtime;
 979         while (offset <= istate->mmap_size - 20 - 8) {
 980                 /* After an array of active_nr index entries,
 981                  * there can be arbitrary number of extended
 982                  * sections, each of which is prefixed with
 983                  * extension name (4-byte) and section length
 984                  * in 4-byte network byte order.
 985                  */
 986                 unsigned long extsize;
 987                 memcpy(&extsize, (char *)(istate->mmap) + offset + 4, 4);
 988                 extsize = ntohl(extsize);
 989                 if (read_index_extension(istate,
 990                                          ((const char *) (istate->mmap)) + offset,
 991                                          (char *) (istate->mmap) + offset + 8,
 992                                          extsize) < 0)
 993                         goto unmap;
 994                 offset += 8;
 995                 offset += extsize;
 996         }
 997         return istate->cache_nr;
 998
 999 unmap:
1000         munmap(istate->mmap, istate->mmap_size);
1001         errno = EINVAL;
1002         die("index file corrupt");
1003 }
1004
1005 int discard_index(struct index_state *istate)
1006 {
1007         int ret;
1008
1009         istate->cache_nr = 0;
1010         istate->cache_changed = 0;
1011         istate->timestamp = 0;
1012         cache_tree_free(&(istate->cache_tree));
1013         if (istate->mmap == NULL)
1014                 return 0;
1015         ret = munmap(istate->mmap, istate->mmap_size);
1016         istate->mmap = NULL;
1017         istate->mmap_size = 0;
1018
1019         /* no need to throw away allocated active_cache */
1020         return ret;
1021 }
1022
1023 #define WRITE_BUFFER_SIZE 8192
1024 static unsigned char write_buffer[WRITE_BUFFER_SIZE];
1025 static unsigned long write_buffer_len;
1026
1027 static int ce_write_flush(SHA_CTX *context, int fd)
1028 {
1029         unsigned int buffered = write_buffer_len;
1030         if (buffered) {
1031                 SHA1_Update(context, write_buffer, buffered);
1032                 if (write_in_full(fd, write_buffer, buffered) != buffered)
1033                         return -1;
1034                 write_buffer_len = 0;
1035         }
1036         return 0;
1037 }
1038
1039 static int ce_write(SHA_CTX *context, int fd, void *data, unsigned int len)
1040 {
1041         while (len) {
1042                 unsigned int buffered = write_buffer_len;
1043                 unsigned int partial = WRITE_BUFFER_SIZE - buffered;
1044                 if (partial > len)
1045                         partial = len;
1046                 memcpy(write_buffer + buffered, data, partial);
1047                 buffered += partial;
1048                 if (buffered == WRITE_BUFFER_SIZE) {
1049                         write_buffer_len = buffered;
1050                         if (ce_write_flush(context, fd))
1051                                 return -1;
1052                         buffered = 0;
1053                 }
1054                 write_buffer_len = buffered;
1055                 len -= partial;
1056                 data = (char *) data + partial;
1057         }
1058         return 0;
1059 }
1060
1061 static int write_index_ext_header(SHA_CTX *context, int fd,
1062                                   unsigned int ext, unsigned int sz)
1063 {
1064         ext = htonl(ext);
1065         sz = htonl(sz);
1066         return ((ce_write(context, fd, &ext, 4) < 0) ||
1067                 (ce_write(context, fd, &sz, 4) < 0)) ? -1 : 0;
1068 }
1069
1070 static int ce_flush(SHA_CTX *context, int fd)
1071 {
1072         unsigned int left = write_buffer_len;
1073
1074         if (left) {
1075                 write_buffer_len = 0;
1076                 SHA1_Update(context, write_buffer, left);
1077         }
1078
1079         /* Flush first if not enough space for SHA1 signature */
1080         if (left + 20 > WRITE_BUFFER_SIZE) {
1081                 if (write_in_full(fd, write_buffer, left) != left)
1082                         return -1;
1083                 left = 0;
1084         }
1085
1086         /* Append the SHA1 signature at the end */
1087         SHA1_Final(write_buffer + left, context);
1088         left += 20;
1089         return (write_in_full(fd, write_buffer, left) != left) ? -1 : 0;
1090 }
1091
1092 static void ce_smudge_racily_clean_entry(struct cache_entry *ce)
1093 {
1094         /*
1095          * The only thing we care about in this function is to smudge the
1096          * falsely clean entry due to touch-update-touch race, so we leave
1097          * everything else as they are.  We are called for entries whose
1098          * ce_mtime match the index file mtime.
1099          */
1100         struct stat st;
1101
1102         if (lstat(ce->name, &st) < 0)
1103                 return;
1104         if (ce_match_stat_basic(ce, &st))
1105                 return;
1106         if (ce_modified_check_fs(ce, &st)) {
1107                 /* This is "racily clean"; smudge it.  Note that this
1108                  * is a tricky code.  At first glance, it may appear
1109                  * that it can break with this sequence:
1110                  *
1111                  * $ echo xyzzy >frotz
1112                  * $ git-update-index --add frotz
1113                  * $ : >frotz
1114                  * $ sleep 3
1115                  * $ echo filfre >nitfol
1116                  * $ git-update-index --add nitfol
1117                  *
1118                  * but it does not.  When the second update-index runs,
1119                  * it notices that the entry "frotz" has the same timestamp
1120                  * as index, and if we were to smudge it by resetting its
1121                  * size to zero here, then the object name recorded
1122                  * in index is the 6-byte file but the cached stat information
1123                  * becomes zero --- which would then match what we would
1124                  * obtain from the filesystem next time we stat("frotz").
1125                  *
1126                  * However, the second update-index, before calling
1127                  * this function, notices that the cached size is 6
1128                  * bytes and what is on the filesystem is an empty
1129                  * file, and never calls us, so the cached size information
1130                  * for "frotz" stays 6 which does not match the filesystem.
1131                  */
1132                 ce->ce_size = htonl(0);
1133         }
1134 }
1135
1136 int write_index(struct index_state *istate, int newfd)
1137 {
1138         SHA_CTX c;
1139         struct cache_header hdr;
1140         int i, removed;
1141         struct cache_entry **cache = istate->cache;
1142         int entries = istate->cache_nr;
1143
1144         for (i = removed = 0; i < entries; i++)
1145                 if (!cache[i]->ce_mode)
1146                         removed++;
1147
1148         hdr.hdr_signature = htonl(CACHE_SIGNATURE);
1149         hdr.hdr_version = htonl(2);
1150         hdr.hdr_entries = htonl(entries - removed);
1151
1152         SHA1_Init(&c);
1153         if (ce_write(&c, newfd, &hdr, sizeof(hdr)) < 0)
1154                 return -1;
1155
1156         for (i = 0; i < entries; i++) {
1157                 struct cache_entry *ce = cache[i];
1158                 if (!ce->ce_mode)
1159                         continue;
1160                 if (istate->timestamp &&
1161                     istate->timestamp <= ntohl(ce->ce_mtime.sec))
1162                         ce_smudge_racily_clean_entry(ce);
1163                 if (ce_write(&c, newfd, ce, ce_size(ce)) < 0)
1164                         return -1;
1165         }
1166
1167         /* Write extension data here */
1168         if (istate->cache_tree) {
1169                 unsigned long sz;
1170                 void *data = cache_tree_write(istate->cache_tree, &sz);
1171                 if (data &&
1172                     !write_index_ext_header(&c, newfd, CACHE_EXT_TREE, sz) &&
1173                     !ce_write(&c, newfd, data, sz))
1174                         free(data);
1175                 else {
1176                         free(data);
1177                         return -1;
1178                 }
1179         }
1180         return ce_flush(&c, newfd);
1181 }